Instructions to use bartowski/stable-code-instruct-3b-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use bartowski/stable-code-instruct-3b-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="bartowski/stable-code-instruct-3b-GGUF")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("bartowski/stable-code-instruct-3b-GGUF", dtype="auto")

llama-cpp-python

How to use bartowski/stable-code-instruct-3b-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="bartowski/stable-code-instruct-3b-GGUF",
	filename="stable-code-instruct-3b-IQ3_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use bartowski/stable-code-instruct-3b-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf bartowski/stable-code-instruct-3b-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf bartowski/stable-code-instruct-3b-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf bartowski/stable-code-instruct-3b-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf bartowski/stable-code-instruct-3b-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf bartowski/stable-code-instruct-3b-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf bartowski/stable-code-instruct-3b-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf bartowski/stable-code-instruct-3b-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf bartowski/stable-code-instruct-3b-GGUF:Q4_K_M

Use Docker

docker model run hf.co/bartowski/stable-code-instruct-3b-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use bartowski/stable-code-instruct-3b-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "bartowski/stable-code-instruct-3b-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "bartowski/stable-code-instruct-3b-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/bartowski/stable-code-instruct-3b-GGUF:Q4_K_M

SGLang

How to use bartowski/stable-code-instruct-3b-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "bartowski/stable-code-instruct-3b-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "bartowski/stable-code-instruct-3b-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "bartowski/stable-code-instruct-3b-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "bartowski/stable-code-instruct-3b-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use bartowski/stable-code-instruct-3b-GGUF with Ollama:
```
ollama run hf.co/bartowski/stable-code-instruct-3b-GGUF:Q4_K_M
```

Unsloth Studio new

How to use bartowski/stable-code-instruct-3b-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for bartowski/stable-code-instruct-3b-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for bartowski/stable-code-instruct-3b-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for bartowski/stable-code-instruct-3b-GGUF to start chatting

Docker Model Runner
How to use bartowski/stable-code-instruct-3b-GGUF with Docker Model Runner:
```
docker model run hf.co/bartowski/stable-code-instruct-3b-GGUF:Q4_K_M
```

Lemonade

How to use bartowski/stable-code-instruct-3b-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull bartowski/stable-code-instruct-3b-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.stable-code-instruct-3b-GGUF-Q4_K_M

List all available models

lemonade list

Llamacpp Quantizations of stable-code-instruct-3b

Using llama.cpp release b2440 for quantization.

Original model: https://huggingface.co/stabilityai/stable-code-instruct-3b

Download a file (not the whole branch) from below:

Filename	Quant type	File Size	Description
stable-code-instruct-3b-Q8_0.gguf	Q8_0	2.97GB	Extremely high quality, generally unneeded but max available quant.
stable-code-instruct-3b-Q6_K.gguf	Q6_K	2.29GB	Very high quality, near perfect, recommended.
stable-code-instruct-3b-Q5_K_M.gguf	Q5_K_M	1.99GB	High quality, very usable.
stable-code-instruct-3b-Q5_K_S.gguf	Q5_K_S	1.94GB	High quality, very usable.
stable-code-instruct-3b-Q5_0.gguf	Q5_0	1.94GB	High quality, older format, generally not recommended.
stable-code-instruct-3b-Q4_K_M.gguf	Q4_K_M	1.70GB	Good quality, similar to 4.25 bpw.
stable-code-instruct-3b-Q4_K_S.gguf	Q4_K_S	1.62GB	Slightly lower quality with small space savings.
stable-code-instruct-3b-IQ4_NL.gguf	IQ4_NL	1.61GB	Good quality, similar to Q4_K_S, new method of quanting,
stable-code-instruct-3b-IQ4_XS.gguf	IQ4_XS	1.53GB	Decent quality, new method with similar performance to Q4.
stable-code-instruct-3b-Q4_0.gguf	Q4_0	1.60GB	Decent quality, older format, generally not recommended.
stable-code-instruct-3b-IQ3_M.gguf	IQ3_M	1.31GB	Medium-low quality, new method with decent performance.
stable-code-instruct-3b-IQ3_S.gguf	IQ3_S	1.25GB	Lower quality, new method with decent performance, recommended over Q3 quants.
stable-code-instruct-3b-Q3_K_L.gguf	Q3_K_L	1.50GB	Lower quality but usable, good for low RAM availability.
stable-code-instruct-3b-Q3_K_M.gguf	Q3_K_M	1.39GB	Even lower quality.
stable-code-instruct-3b-Q3_K_S.gguf	Q3_K_S	1.25GB	Low quality, not recommended.
stable-code-instruct-3b-Q2_K.gguf	Q2_K	1.08GB	Extremely low quality, not recommended.

Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski

Downloads last month: 593

GGUF

Model size

3B params

Architecture

stablelm

Hardware compatibility

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

Evaluation results

pass@1 on MultiPL-HumanEval (Python)
self-reported

32.400
pass@1 on MultiPL-HumanEval (C++)
self-reported

30.900
pass@1 on MultiPL-HumanEval (Java)
self-reported

32.100
pass@1 on MultiPL-HumanEval (JavaScript)
self-reported

32.100
pass@1 on MultiPL-HumanEval (PHP)
self-reported

24.200
pass@1 on MultiPL-HumanEval (Rust)
self-reported

23.000